This research is a continuation of our efforts to build an automatic microscope system using computer pattern recognition methods to, in the first instance, perform a differential white blood cell count by machine. The completion of the hardware and an evaluation of the performance of all of its components using the previously developed software is the first objective. This will be followed by more detailed R&D on selected aspects of the system to improve its performance, such as the nuclear isolation problem, a hierarchical classification scheme and a faster and more accurate cell finder. Extensions of automatic cytology procedures to other problems, such as the differential classification of leukemia will be investigated. The results we expect to obtain on various phases of such a system are expected to be useful to prospective users of commercial equipment as well as to manufacturers who need basic facts from which to design automated equipment. Such equipment is expected to help provide the technological component of the health services for the future.